P reve ntin g D i ab et i c

Ketoacidosis
Craig A. Jefferies, MBChB, MDa,b,*,1, Meranda Nakhla, MSc, MD, FRCPCc,1,
Jos G.B. Derraik, PhDb, Alistair J. Gunn, MBChB, PhDa,d,
Denis Daneman, MBBCh, FRCPC, DSc (Med)e, Wayne S. Cutfield, MBChB, MDa,b
KEYWORDS
 Type 1 diabetes mellitus  Diabetic ketoacidosis  Prevention  Education
 Recurrent  Noncompliance  Girls
KEY POINTS
 Diabetic ketoacidosis (DKA) is a complication of severe insulin deficiency leading to hyperglycemia, with its attendant glycosuria, dehydration, and ketogenesis, leading to
acidosis.
 DKA is associated with significant morbidity, non-negligible mortality, and excessive
health care expenditures.
 DKA is largely an avoidable and preventable complication of type 1 diabetes mellitus
(T1DM).
 Risk factors associated with DKA at diagnosis of T1DM include age less than 5 years old,
lack of private health insurance, lower socioeconomic status (SES), misdiagnosis, delayed
diagnosis, and living in regions with a low background incidence of T1DM.
 Factors associated with increased DKA risk in children with established T1DM include age
greater than 13 years, female gender, low SES, lack of private health insurance, poor
family functioning, psychiatric disorders, higher reported insulin dose, and poor glycemic
control.

Competing Interests: The authors have no financial or nonfinancial conflicts of interest to

disclose that may be relevant to this work.
a
Paediatric Endocrinology Service, Starship Childrens Hospital, Auckland District Health
Board, 2 Park Road, Auckland, 1023, New Zealand; b Liggins Institute, University of Auckland,
85 Park Road, Auckland, 1023, New Zealand; c Department of Paediatrics, Montreal Childrens
Hospital, McGill University, 2300 Tupper Street, H3H 1P3, Montreal, Canada; d Department of
Physiology, University of Auckland, 85 Park Road, Auckland, 1023, New Zealand;
e
Department of Pediatrics, The Hospital for Sick Children, University of Toronto, 555 University
Avenue, M5G 1X8, Toronto, Canada
1
Coprimary authors.
* Corresponding author. Starship Childrens Health, Private Bag 92024, Auckland 1142, New
Zealand.
E-mail address: CraigJ@adhb.govt.nz
Pediatr Clin N Am 62 (2015) 857871
http://dx.doi.org/10.1016/j.pcl.2015.04.002
pediatric.theclinics.com
0031-3955/15/$ see front matter 2015 Elsevier Inc. All rights reserved.

858

Jefferies et al

INTRODUCTION

T1DM is one of the most common chronic diseases of childhood with significant
morbidity and non-negligible mortality.1,2 The global incidence of T1DM is increasing
by 3% per year in children and adolescents and by 5% per year in preschoolers.3,4
Acute complications, such as DKA, are potentially avoidable causes of hospitalizations
and mortality in children and adolescents with T1DM. DKA occurs in the presence of
severe insulinopenia and is characterized by hyperglycemia, acidosis, and ketosis.
The early development of T1DM involves progressive insulin deficiency, leading to
weeks of symptomatic hyperglycemia with polyuria, polydipsia, and weight loss.5,6
If left untreated, the combination of insulin deficiency and stress (mediated through
increased circulating levels of counter-regulatory hormones, including cortisol, catecholamines, growth hormone, and glucagon) leads to lipolysis. Hepatic metabolism
of free fatty acids as an alternative energy source (ie, ketogenesis) results in accumulation of acidic intermediate and end metabolites (ie, ketones and ketoacids). The predominant ketones in the body are b-hydroxybutyrate, acetate, and acetoacetate.5,6 In
DKA, the ratio of b-hydroxybutyrate to acetoacetate increases from equimolar
amounts (1:1) up to 1.3:1 to 5.5:1.7 Accumulation of ketoacids is due to their ratelimiting metabolism. Diagnosis of DKA has been traditionally made with a combination
of hyperglycemia (blood glucose >11 mmol/L or 200 mg/dL), venous pH less than 7.3
and bicarbonate levels less than 15 mmol/L.8
DKA at initial presentation is common, but rates vary strikingly between countries,
ranging from approximately 13% to 80%.911 The highest rates are typically seen in
the developing world, not surprisingly given the resource constraints.11 DKA in children previously diagnosed with T1DM is also unfortunately common, with highly variable rates between centers and countries.11,12 DKA remains the most common cause
of mortality in T1DM. The most devastating consequence of DKA is cerebral edema
with an incidence of 0.5% to 0.9%.13,14 The mortality rate from cerebral edema is
21% to 24%, and morbidity from serious neurologic sequelae occurs in 15% to
35% of cases.11,12 Furthermore, DKA has a long-term impact on cognitive function,
including a fall in IQ and persistent loss of short-term memory. The health care expenditures associated with 1 episode of DKA are substantial, estimated at more than
$3500.15
EPIDEMIOLOGY AND CAUSES OF DIABETIC KETOACIDOSIS

DKA occurs generally under 2 circumstances and in a limited number of situations:

1. At diagnosis
2. In those with established diabetes
a. As a result of accidental or deliberate insulin omission or
b. Inappropriate management of intercurrent illness
Most, if not all, episodes of DKA can be prevented. These episodes generally represent a failure of timely diagnosis and/or management of T1DM. Understanding the
epidemiology and factors associated with the onset of DKA is essential in developing
measures to prevent these episodes.
DIABETIC KETOACIDOSIS AT DIAGNOSIS

Symptoms of new-onset T1DM typically develop over a few days to weeks before presentation, whereas DKA rapidly evolves once ketones accumulate16: 75% of patients
with new cases of T1DM in the EURODIAB (the Epidemiology and Prevention of

Preventing Diabetic Ketoacidosis

Diabetes) project had symptoms for more than 2 weeks before diagnosis.17 Thus, DKA
at diagnosis of new T1DM ultimately reflects lack of awareness by parents and primary
care physicians of the evolving symptoms of T1DM.17
A systematic review in 2011 identified 46 studies, including more than 24,000 children with T1DM from 31 countries18: the frequency of DKA at diagnosis varied 6-fold,
from 12.8% to 80% (Table 1).16 Rates were lowest in Sweden, and Canada and highest in the United Arab Emirates, Saudi Arabia, and Romania. There was an inverse correlation between DKA frequency at diagnosis and the regional background incidence
of T1DM, findings that are consistent with previous studies from Europe. This may be
explained by an overall increased awareness of T1DM and its presenting symptoms,
leading to earlier recognition and diagnosis. DKA frequencies were also found to be
lower in countries further from the equator and with a higher gross domestic product
(GDP).12 There are several explanations for this association. Latitude may represent a
group of characteristics, including economy, health care provision, and disease
burden.12 Usher-Smith and colleagues12 found a colinearity between latitude and a
countrys GDP, suggesting that the variation in latitude may be explained more by a
countrys economy and health care provision than distance from the equator. In
addition, in a recent report analyzing data from countries identified as wealthy by
the Organisation for Economic Co-operation and Development, a close correlation
was found between DKA risk at disease onset and income inequality, defined as the
difference in average incomes between a nations highest and lowest income
earners.19 Among the countries included in the analysis, the frequency of DKA at diagnosis ranged from 16% to 54%, with countries with the lowest income inequality having lower DKA frequencies.
Table 1
Frequency of diabetic ketoacidosis at disease onset of type 1 diabetes mellitus
Country

Diabetic Ketoacidosis (%)

United Arab Emirates

80

Romania

67

Saudi Arabia

4077

France

54

Kuwait

3849

Poland

3354

China

42

Oman

42

Lithuania

35

Austria

37

Italy

3241

Bulgaria

35

Germany

2653

United States

2744

Turkey

29

United Kingdom

2538

Ireland

25

Finland

1922

Canada

18.6

Sweden

1314

859

860

Jefferies et al

The key individual factors associated with greater risk of DKA were being less than
2 years old at presentation, being incorrectly diagnosed or having treatment delayed,
belonging to an ethnic minority, lower SES, lack of health insurance in the United
States, lower parental education, lower body mass index, and preceding infection.20
Conversely, having a first-degree relative with T1DM was protective (odds ratio
[OR] 0.33).
A small qualitative study (16 children) within the East of England found that the families were aware of symptoms for several weeks before a trigger, such as weight loss,
led them to seek medical advice.21 At the time of presentation, many families suspected the diagnosis. A strong protective effect of a family history of T1DM has
been noted in several studies.2123 This relationship is unlikely to be genetic, because
although high-risk genotypes are associated with DKA risk, having even 1 first-degree
family member with T1DM is associated with dramatically reduced risk of DKA.24
Furthermore, participating in a long-term follow-up study, The Environmental Determinants of Diabetes in the Young, was associated with reduced risk of DKA.25 Thus, this
protective relationship is almost certainly due to greater family awareness of the
symptoms of diabetes.
Supporting this proposal, studies have found an overall fall in the risk of DKA over a
20-year interval in northern Finland,26 consistent with the overall increase in incidence
of T1DM over this period. There are some contradictory findings, however. In some
countries, the rate of new-onset DKA has not fallen over time despite a temporal increase in incidence, even where there are highly organized and easily accessible
health care systems. In Austria for example, the risk of DKA on a prospectively
recorded country-wide register remained similar between 2005 to 2009 and 2010 to
2011.27 A single-center study from Australia found no change in the risk of DKA in
1073 children and adolescents ages up to 18 years with newly diagnosed T1DM
from 1998 to 2010.28 Similarly, in Auckland, the largest city in New Zealand, the incidence and severity of new-onset DKA remained stable at approximately 27% over the
past 15 years29 and are similar to the previous review in 1995 to 1996.10 These findings
suggest that health care providers and families are failing to recognize the early symptoms of T1DM.
Age

Several studies have consistently demonstrated that younger age at diagnosis of

T1DM is a significant risk factor in the development of DKA. Usher-Smith and
colleagues18 conducted a meta-analysis consisting of 32 studies and found that children ages less than 2 years old had 3 times the risk of presenting with DKA as children
ages greater than 2 years old (OR 3.41; 95% CI, 2.544.59). This increased risk
continued up to age 5 years (OR 1.59; 95% CI, 1.381.84). In general, studies have
demonstrated that the frequency of DKA at diagnosis decreases significantly with
age from 37% to 40% (95% CI, 32.9%41.8%) in children ages 0 to 4 years old to
14.7% to 23.6% (95% CI, 11.7%17.7%) in children ages 15 to 19 years old.23,30
The higher rates of DKA in the younger age groups may be related to several factors,
including (1) the classic symptoms of T1DM may not be obvious and easily distinguishable from other common acute illnesses, resulting in a delay in diagnosis; (2) clinicians may have a lower index of suspicion for T1DM among younger children,
particularly in the under 2-year age group;3,25 (3) younger children have a less developed mechanism of metabolic compensation resulting in faster development of
acidosis and dehydration;23,31 and (4) b-cell destruction4 may be more aggressive in
younger children because serum levels of C-peptide are lowest in the under 2-year
age group at diagnosis of T1DM compared with older age groups.32

Preventing Diabetic Ketoacidosis

Ethnic Minority

Studies in both the United States and United Kingdom have demonstrated an
increased risk of DKA among ethnic minorities. A recent multicenter US study found
that nonwhite youth were more at risk of presenting with DKA at diagnosis compared
with non-Hispanic white youth, independent of SES and age (OR 1.21; 95% CI, 1.03
1.43).20 A study in the United Kingdom reported a higher frequency of DKA at diagnosis among ethnic minorities (41.3%) compared with non-Hispanic white patients
(21.4%) (P 5 .03).33 In addition, a study in New Zealand showed that non-Europeans
were more likely to present in DKA than those of European ethnicity (OR 1.52; P 5
.048).29 These disparities may be due to cultural or language barriers, lack of access
to health care services, or lack of awareness of T1DM in ethnic minorities.
Lower Socioeconomic Status

Lower SES as measured by family income or neighborhood income is associated with

an increased risk of DKA at diagnosis.34 One US study found that the frequency of DKA
was 70% to 80% higher in patients with a family income below $50,000 compared with
those with a family income greater than $50,000 (higher income).20 In Canada, where
patients enjoy universal access to health care, patients with low SES, as measured by
neighborhood income quintiles, were more likely to present with DKA at diagnosis
compared with the highest income quintiles (OR 1.39; 95% CI, 1.17-1.63).34
Lack of Private Health Insurance

In countries without universal health coverage, lack of private health insurance has
been consistently found to be a significant risk factor for DKA at diagnosis.35,36 For
instance, study in the United States found that patients with private health insurance
were least likely to present with DKA.35 Patients with no health insurance were 60%
more likely to present with DKA compared with those with health insurance coverage.
Furthermore, when uninsured patients presented with DKA, they were more likely to
present with severe DKA.37 The odds of uninsured patients presenting with severe
DKA (ie, venous pH <7.10; serum bicarbonate <5 mmol/L) were 6 times (OR 6.09;
95% CI, 3.2111.56) the odds for insured patients.36 These findings suggest that patients without health insurance may delay seeking timely medical care and thus present with severe DKA.
Delayed Diagnosis

A delay of more than 24 hours between the initial presentation to a primary or secondary
care provider and referral to a multidisciplinary diabetes team in the United Kingdom
has been reported associated with an increased risk of presenting with DKA (52.3%
vs 20.5%, P<.05).33 The most common reason cited for the delayed diagnosis was
arrangement for a fasting blood sugar prior to referral to a multidisciplinary team.
Diagnostic Error

Several retrospective cohort studies have reported an increased risk of DKA in children
not diagnosed with T1DM at first presentation to the health care system due to either a
misdiagnosis or lack of recognition of the symptoms of T1DM.33,34 In Ontario, Canada,
the authors previously reported that children with DKA at diabetes diagnosis were
more likely to have seen a health care professional on 1 or more occasions in the weeks
prior, at which time the diagnosis of T1DM was missed.34 Diagnostic error at first presentation to the health care system has been associated with a 3-fold increased risk of
presenting with DKA (OR 3.35; 95% CI, 2.35-5.79), independent of the presence of a

861

862

Jefferies et al

preceding infectious illness.1 Misdiagnosis was more likely to occur among younger
children. In 1 study, patients with DKA in whom the diagnosis of diabetes was missed
at initial presentation tended to be younger (mean age 5.4  4.4 years old) compared
with those in whom the diagnosis was not missed (mean age 8.8  4.0 years old)
(P<.001). Younger children tended to be misdiagnosed with urinary tract infection, upper respiratory tract infection, gastroenteritis, or otitis media.34
Preceding Infectious Illness

The presence of a concomitant infection may mask the early symptoms of diabetes or
delaying the diagnosis resulting in an increased risk of DKA.20 Furthermore, an infectious or febrile illness generally increases the counter-regulatory response, resulting in
insulin resistance and metabolic decompensation.
Protective Factors

The protective factors associated with a decreased risk of DKA at disease onset
include first-degree relative with T1DM and higher parental education. Having a higher
awareness and better recognition of the signs and symptoms of hyperglycemia likely
explain the decreased risk of DKA among families with a first-degree relative with
diabetes.1,30,32 Children from families in which parents had greater than a postsecondary education or had at least 1 parent with an academic degree are at a decreased
risk of DKA at diagnosis.18,30,32
Can Diabetic Ketoacidosis Be Prevented at Diagnosis?

In light of the existing evidence, it is highly plausible that improving community and
medical awareness of the presenting features of T1DM in children may help prevent
DKA. There have been few studies, however, on community education with indirect
and unmonitored methods used, which have led to inconclusive results.
A small study from Parma, Italy,38 involved providing a poster to primary and secondary public schools and providing local pediatricians with equipment for measuring
urine and blood glucose and patient cards. Over the subsequent 8 years, the rate of
DKA in children ages 6 to 14 years was 12.5% (3/24) in Parma compared with 83%
(25/30) in nearby areas, where children did not receive the information. Furthermore,
the duration of symptoms was much shorter in the Parma area (mean 5 vs 28 days),
and the length of hospitalization was also shortened. Despite its encouraging results,
this study was limited by its small size and the high baseline rate of DKA compared
with typical European rates.
Another small prospective study also suggested significant benefit from an educational campaign in Australia. In this study, a 2-year control period was followed by a
2-year intervention37 in 1 region (Gosford) but not in 2 others (Newcastle and Sydney).
Posters were provided to childcare centers, schools, and doctors offices, with doctors also given glucose and ketone testing equipment. The proportion of children
with DKA decreased in the intervention region from 38% (15/40) to 14% (4/29),
whereas there was no change in DKA rates in the control regions (37% [46/123] and
39% [49/127]). Unexpectedly, the incidence of T1DM in Gosford also fell by 28%,
raising the possibility that the reduction in new-onset DKA might have partly reflected
redirection of patients to other larger, childrens hospitals in the region.
In contrast, a national study in Wales found no benefit.39 Posters were sent to every
pharmacy, school, and general practitioner surgery across Wales, and the campaign
publicized with radio interviews. Before the campaign, the risk of DKA was approximately 25% from 1991 to 2009, with no temporal change. After the campaign, from
2008 to 2009, 26% of children diagnosed with T1DM had DKA. Questionnaires

Preventing Diabetic Ketoacidosis

suggested that few families were aware of the campaign. Thus, it seems that the
format of the campaign without more direct engagement with the community may
not be effective.
Furthermore, another nationwide prospective study in Austria involving 4038 children with new-onset T1DM also found no effect of community education.27 This study
attempted greater penetration than other studies, and posters were given to all kindergartens, schools, pharmacies, pediatricians, and general practitioners in 2009. In
addition, medical officers in schools received education twice a year, and there was
national media publicity. Before the campaign (20052009), 26% of children had
mild DKA and 12% had severe DKA compared with 27% mild and 9.5% severe cases
afterward (20102011). Although DKA was more common in younger children, there
was no effect of stratification by age. This study suggests that a primarily posterbased education campaign is insufficient, possibly because of greater competition
for parents attention. Regional differences cannot be excluded, however, and it is
also possible that simply displaying posters was the wrong media tool for the modern
generation.
Thus, there is at best limited evidence that community-based mass campaigns
improve the risk of DKA in new-onset T1DM. Critical to any mass educational program
is a clear simple message. It is worth reflecting that in the Parma study, approximately
90% of children had bed wetting at diagnosis.38 Similarly, in Auckland, a local audit
from 2000 to 2009 suggests that more than 95% of children newly diagnosed with
T1DM had bed wetting or increased night-time urination (Starship Childrens Hospital,
unpublished data). The authors speculate that a focus on such a simplified message
may be more effective for community education and should be tested.
DIABETIC KETOACIDOSIS IN ESTABLISHED TYPE 1 DIABETES MELLITUS

Diabetes mellitus management in children is complex and requires intense resources

as well as regular access to timely and effective health care services to prevent complications. T1DM is an ambulatory-care sensitive condition, defined as a group of
medical conditions typically managed in an ambulatory setting, whereby an admission
to hospital indicates a potentially avoidable complication and is an indirect indicator of
access and overall quality of outpatient care.40,41 DKA is generally avoidable and its
occurrence in children and adolescence with established T1DM may indicate poor
outpatient diabetes mellitus care and could be prevented with comprehensive multidisciplinary outpatient care and improved adherence to diabetes management.
Table 2 outlines the risk factors for DKA among youth with established T1DM.
Epidemiology

Globally, reported rates of DKA in children with established T1DM range between 1
per 100 patient-years in a Swedish cohort35 to 12 episodes per 100 patient-years in
US cohort.36,42 A retrospective cohort study conducted in the United States by
Rewers and colleagues36 determined that the incidence of DKA was 8 per 100
patient-years in 1243 children from 1996 to 2000. This incidence increased in age
among girls (4 per 100 patient-years in <7 year olds; 8 in 712 year olds; and 12
in 13 year olds) but not boys. In addition, approximately 60% of the DKA episodes
occurred in 5% of children with recurrent events (>2 episodes). In a more recent retrospective cohort study from Europe (19952008), the incidence of DKA among 28,770
patients with T1DM was 6.3 per 100 patient-years and remained relatively unchanged
over the study period, with recurrent episodes seen in 1% of patients.43 The overall
incidence of DKA was slightly higher in girls (7% vs 5.8%) and in immigrants, but there

863

864

Jefferies et al

Table 2
Risk factors associated with new-onset diabetic ketoacidosis versus recurrent diabetic
ketoacidosis
New-onset Diabetic Ketoacidosis

Recurrent Diabetic Ketoacidosis

Young age (especially <2 y)

Older age (especially adolescents)

Lack of access to medical care

Lack of access to medical care

Lack of awareness of diabetes

Psychiatric disorders

Lack of family history of T1DM

Low SES

Delayed medical diagnosis of T1DM

Low family cohesion

Low parental education

Higher HbA1c

Recent infection

Low income

Minority ethnic groups

Minority ethnic groups

High insulin doses and noncompliance with

insulin therapy

was no effect of treatment type or duration of diabetes. Recurrent DKA was associated with older age (in particular the early teenage years), higher hemoglobin A1c
(HbA1c) levels, and higher insulin doses.
In the United States, in the T1D Exchange Clinic Registry of 13,487 children and
young adults (<26 years of age) with T1DM for at least 2 years,35 DKA was most
frequent in adolescents and associated with higher HbA1c, nonwhite race, lack of private health insurance, and lower household income. A small Swedish study identified
142 episodes of DKA in 1990 to 2000; subjects were aged 14.6 years on average with a
mean T1DM duration of 6.6 years.44 The reported triggers of DKA included missed insulin doses (48.6%), gastroenteritis (14.1%), technical pump problems (12.7%), infections (13.4%), and social problems (1.4%). Pump users had approximately double the
rate of DKA compared with patients using intermittent injections. This finding is not
universal, however, and may reflect an improved understanding of insulin pumping
and early testing for blood ketones.44 Although the risk of DKA increases with duration
of diabetes, there is an appreciable rate in the first year after diagnosis. In a
population-based study in Germany, 373 subjects with newly diagnosed T1DM
aged less than 15 years who were followed for a mean of 1 year had 4.7 times higher
risk of hospitalization.45
Risk Factors in Children with Established Type 1 Diabetes Mellitus

In a prospective cohort study of 1243 children and youth with T1DM, the risk of DKA
increased with higher HbA1c and higher reported insulin doses among children ages
less than 13 years old (see Table 2).45,46 Among older children (>13 years old), the
risk of DKA increased with higher HbA1c (relative risk [RR] 1.43 per 1% increase;
95% CI, 1.31.6), higher reported insulin dose, underinsurance, and the presence of
psychiatric disorders, including major depression, bipolar or anxiety disorder, and/
or use of psychotropic medications. These factors were also found to confer an
increased risk of recurrent DKA. Other studies have also found similar risk factors
for DKA.35,42,45
The higher risk of DKA among adolescent girls may be related to body image issues,
whereby 10% of adolescent girls with T1DM meet Diagnostic and Statistical Manual of
Mental Disorders47 criteria for eating disorders compared with 4% of their agematched controls without diabetes.48 One aspect of T1DM that contributes to the
development of an eating disorder is that adolescent girls may deliberately omit insulin

Preventing Diabetic Ketoacidosis

as a unique and readily available way to control weight through induced hyperglycemia and glycosuria, increasing the risk of DKA.4951
Higher reported insulin dose may represent either lower endogenous insulin secretion with longer duration of T1DM or insulin resistance related to puberty or obesity.45
Alternatively, in patients who omit insulin, a higher prescribed insulin dose may not be
the actual dose taken.
Psychiatric conditions are a significant risk factor for DKA, particularly among
adolescent girls (RR 3.22; 95% CI, 2.254.61).8,52 This association may be due to
nonadherence to diabetes self-management and to insulin omission.53
Lower SES as measured by lack of private health insurance, head-of-household unemployment, or low household income has consistently been found associated with
an increased risk of DKA.36 This association maybe driven by factors, such as compliance with medical care, including decreased blood glucose monitoring, access to
transportation, and irregular physician visits.
Additional risk factors include longer diabetes duration,54 family functioning, the
transition to adult care,55 and insulin pump therapy.46 Several studies suggest that
family functioning is associated with DKA risk. In 1 study, children who had an
episode of DKA were more likely to perceive their parents as less warm and caring
toward their diabetes management (measure of diabetes-specific family support) and
had caregivers who rated themselves as less supportive of their childs diabetes care
(measure of parental negativity) compared with children who did not have a DKA
episode.55
The transition to adult care occurs during a critical and vulnerable period for those
with T1DM. The authors data from the province of Ontario, Canada, identified a small
but significant increase in DKA rates as youth with T1DM transitioned from a pediatric
to adult health care team. Furthermore, the authors found that girls and lower SES, as
measured by neighborhood level income, decreased local medical resources, and
previous hospitalizations for DKA, were significant predictors of post-transition DKA
episodes.55
In summary, DKA in children and adolescents with established T1DM is an avoidable complication that can be prevented through appropriate education and psychosocial support, consistent self-monitoring of blood glucose and ketones, and
improved adherence to diabetes self-management.
Poor glycemic control

Poor glycemic control is a major risk factor for recurrent DKA in teenagers.49 In turn,
poor glycemic control is directly related to poor adherence with insulin treatment. For
example, in Scotland, in a cohort of 89 patients at a mean age 16 years, there was a
significant inverse association between HbA1c and adherence (r2 5 0.39).40 Not surprisingly, poor adherence was associated with more admissions for DKA and the
secular trend for deterioration in glycemic control from 10 to 20 years.
Small studies in which children were studied in more detail have highlighted a clear
role for underlying behavioral and psychological stresses in association with recurrent
DKA. In a sample of 92 school-age children followed for an average of 9 years from
diagnosis, 28% had 3 or more admissions for DKA. Recurrent DKA was associated
with higher HbA1c levels, more behavioral problems, younger age at diagnosis, and
lower SES.54 Similarly, in a cohort of 61 children and adolescents followed for 8 years
after diagnosis, 28% had 1 or more episode of DKA.56,57 DKA was more common in
girls and was associated with more behavior problems and lower social competence,
higher levels of reported family conflict, and lower levels of family cohesion, expressiveness, and organization in year 1.

865

866

Jefferies et al

These studies indicated that a mixture of poor background glycemic control, disadvantaged groups, and (particularly in the United States) lack of medical insurance as
key risk factors for recurrent DKA. Nonetheless, even countries with comprehensive
medical systems have found little change in the rate of DKA over time. For example,
in Ontario, Canada, there were 5008 hospital admissions for DKA in children younger
than 19 years from 1991 to 1999, with no change in the admission rate or the case
fatality rate of 0.18% over this time.57
Prevention of recurrent diabetic ketoacidosis

The existing evidence suggests that there are 2 broad targets for preventing recurrent
DKA: improving access to medical care and improving management of individual
patient approach to diabetes. The impact of social and access issues in deprived populations is substantial. For example, among adult urban African Americans who developed DKA, more than two-thirds of cases occurred due to cessation of insulin therapy
because of lack of money for insulin or transportation to the hospital as well as limited
self-care skills.38 A recent study in the same setting again highlighted the strong link
between DKA and lack of adherence to insulin treatment in adult inner-city minority
patients in the United States.58 Once more, two-thirds of cases of DKA were due to
cessation of insulin therapy and were linked to depression, alcohol and drug abuse,
and homelessness (but not psychiatric illnesses). Such high-risk populations would
benefit from practical, targeted assistance.
The more common background for recurrent DKA, however, is teenage patients
from nondisadvantaged populations. A systematic review in 2000 identified 24 randomized controlled trials of behavioral interventions for adolescents with T1DM.53
The effect on DKA rates was rarely examined in the individual studies; therefore,
DKA was not a focus of the review. The effect on glycemic control was small and
notably heterogeneous, equivalent to a reduction in HbA1c of 0.3%. Furthermore,
the studies were small, there were few common outcomes to allow direct comparison
of interventions, and there was no long-term follow-up. Moreover, the most effective
interventions seem to involve such high institutional and patient commitments that it is
questionable whether they could be translated to routine care.
Subsequently, there have been several small studies with similar limitations. For
example, a small study in which 17 children received ten 90-minute sessions of inhome behavioral family systems therapy had no effect on glycemic control.59,60 A
modified version of this program in 104 families (involving either educational support
for multiple families or 12 sessions over 6 months) was associated with improved glycemic control up to 18 months from the start of the program, but its cost-effectiveness
and effect on DKA risk are still unclear. Scheduled bimonthly telephone calls over
17 months from a pediatric diabetes educator did not improve glycemic control or
reduce rates of hospital admission in 123 patients at a mean age of 12 years.61 A
more ambitious program consisted of a 6-month home-based intervention for 37
adolescents with poorly controlled T1DM (HbA1c >9.0%), involving monthly home
visits and weekly phone contact.62 In comparison to 32 adolescents in routine care
only, there was a modest improvement in HbA1c, from 11.1% to 9.7% at 6 months,
that was not maintained at 12 and 18 months. It is unknown whether continuing treatment would have been more effective.
In contrast, recent larger studies of behavioral interventions have yielded encouraging results. In a randomized controlled trial in 127 adolescents with poorly controlled
T1DM, intensive home-based psychotherapy over 2 years was associated with sustained reduction in hospital admissions compared with baseline or to controls over
the full period.41 This intensive intervention was expensive ($6934 per patient), but

Preventing Diabetic Ketoacidosis

this cost was offset by reduced admissions for DKA. A subsequent trial in 146 adolescents with T1DM or type 2 diabetes mellitus showed that this approach was superior
to telephone support in improving glycemic control over 12 months.63 Another study
involved more general training in adolescents before 20 years of age, based on 6 small
group sessions and monthly follow-up, including social problem solving, cognitive
behavior modification, and conflict resolution.64 There was an associated improvement in glycemic control, and the impact of diabetes on their quality of life was
lessened compared with intensive diabetes management alone. This intervention,
however, did not reduce the rate of DKA.
On the other hand, in patients with recurrent DKA, Golden and colleagues65 demonstrated a significant decrease in DKA episodes among patients who were given their
insulin injections by a responsible adult. Similarly, Nguyen and colleagues66 demonstrated a significant improvement in HbA1c levels when teens with persistently poor
metabolic control received a dose of long-acting insulin analog at lunchtime by a
school nurse.
SUMMARY

The factors associated with DKA at presentation and recurrent DKA after diagnosis
are different. Broadly, DKA at presentation seems to be a problem of lack of awareness of the symptoms of T1DM, whereas recurrent DKA in patient with established
diabetes seems primarily related to omission of insulin, augmented by social issues.
In cases of DKA at presentation, despite promising results from 2 community education projects, nationwide educational campaigns have been ineffective. The authors
speculate that simplified messages and better ways of increasing awareness are
needed.
In contrast, recurrent DKA is a problem of a minority, in particular girls in early
adolescence with related family and behavioral issues, often from disadvantaged
backgrounds. The key to reducing recurrent DKA is likely to be achieved by providing
targeted intervention for high-risk patients. There is increasing evidence that several
forms of support can modestly reduce DKA, but the most promising interventions
involve a substantial and expensive commitment and have not been validated in large
trials. The challenge for future research is to find programs that are not only efficacious
but also readily applicable in routine clinical practice.
REFERENCES

1. Edge JA, Ford-Adams ME, Dunger DB. Causes of death in children with insulin
dependent diabetes 1990-96. Arch Dis Child 1999;81(4):31823.
2. Dahlquist G, Kallen B. Mortality in childhood-onset type 1 diabetes: a populationbased study. Diabetes Care 2005;28(10):23847.
3. Patterson CC, Dahlquist GG, Gyurus E, et al. Incidence trends for childhood type
1 diabetes in Europe during 1989-2003 and predicted new cases 2005-20: a
multicentre prospective registration study. Lancet 2009;373(9680):202733.
4. The Diamond Project Group. Incidence and trends of childhood Type 1 diabetes
worldwide 1990-1999. Diabet Med 2006;23(8):85766.
5. Lawrence SE, Cummings EA, Gaboury I, et al. Population-based study of incidence and risk factors for cerebral edema in pediatric diabetic ketoacidosis.
J Pediatr 2005;146(5):68892.
6. Edge JA, Hawkins MM, Winter DL, et al. The risk and outcome of cerebral
oedema developing during diabetic ketoacidosis. Arch Dis Child 2001;85(1):
1622.

867

868

Jefferies et al

7. Wolfsdorf JI. The International Society of Pediatric and Adolescent Diabetes

guidelines for management of diabetic ketoacidosis: Do the guidelines need to
be modified? Pediatr Diabetes 2014;15(4):27786.
8. Wolfsdorf JI, Allgrove J, Craig ME, et al. Diabetic ketoacidosis and hyperglycemic hyperosmolar state. Pediatr Diabetes 2014;15(Suppl 20):15479.
9. Dunger DB, Sperling MA, Acerini CL, et al. ESPE/LWPES consensus statement on
diabetic ketoacidosis in children and adolescents. Arch Dis Child 2004;89:18894.
10. Jackson W, Hofman PL, Robinson EM, et al. The changing presentation of children with newly diagnosed type 1 diabetes mellitus. Pediatr Diabetes 2001;2:
1549.
11. Onyiriuka AN, Ifebi E. Ketoacidosis at diagnosis of type 1 diabetes in children
and adolescents: frequency and clinical characteristics. J Diabetes Metab Disord
2013;12:47.
12. Usher-Smith JA, Thompson M, Ercole A, et al. Variation between countries in the
frequency of diabetic ketoacidosis at first presentation of type 1 diabetes in children: a systematic review. Diabetologia 2012;55(11):287894.
13. Cameron FJ, Scratch SE, Nadebaum C, et al. Neurological consequences of diabetic ketoacidosis at initial presentation of type 1 diabetes in a prospective cohort
study of children. Diabetes Care 2014;37(6):155462.
14. Glaser N, Barnett P, McCaslin I, et al. Risk factors for cerebral edema in children
with diabetic ketoacidosis. N Engl J Med 2001;344(4):2649.
15. Tieder JS, McLeod L, Keren R, et al. Variation in resource use and readmission for
diabetic ketoacidosis in childrens hospitals. Pediatrics 2013;132:22936.
16. Wolfsdorf J, Craig ME, Daneman D, et al. Diabetic ketoacidosis. Pediatr Diabetes
2007;8:2843.
17. Levy-Marchal C, Patterson CC, Green A. Geographical variation of presentation
at diagnosis of type I diabetes in children: the EURODIAB study. Diabetologia
2001;44(Suppl 3):B7580.
18. Usher-Smith JA, Thompson MJ, Sharp SJ, et al. Factors associated with the presence of diabetic ketoacidosis at diagnosis of diabetes in children and young
adults: a systematic review. BMJ 2011;343:d4092.
19. Limenis E, Shulman R, Daneman D. Is the frequency of ketoacidosis at onset of
type 1 diabetes a child health indicator that is related to income inequality? Diabetes Care 2012;35(2):e5.
20. Dabelea D, Rewers A, Stafford JM, et al. Trends in the prevalence of ketoacidosis
at diabetes diagnosis: the SEARCH for diabetes in youth study. Pediatrics 2014;
133(4):e938945.
21. Usher-Smith JA, Thompson MJ, Walter FM. Looking for the needle in the haystack: a qualitative study of the pathway to diagnosis of type 1 diabetes in children. BMJ Open 2013;3:e004068.
22. Choleau C, Maitre J, Filipovic Pierucci A, et al. Ketoacidosis at diagnosis of type
1 diabetes in French children and adolescents. Diabetes Metab 2014;40:
13742.
23. de Vries L, Oren L, Lazar L, et al. Factors associated with diabetic ketoacidosis at
onset of Type 1 diabetes in children and adolescents. Diabet Med 2013;30:13606.
24. Marigliano M, Morandi A, Maschio M, et al. Diabetic ketoacidosis at diagnosis:
role of family history and class II HLA genotypes. Eur J Endocrinol 2013;168:
10711.
25. Elding Larsson H, Vehik K, Bell R, et al. Reduced prevalence of diabetic ketoacidosis at diagnosis of type 1 diabetes in young children participating in longitudinal follow-up. Diabetes Care 2011;34:234752.

Preventing Diabetic Ketoacidosis

26. Hekkala A, Knip M, Veijola R. Ketoacidosis at diagnosis of type 1 diabetes in children in northern Finland: temporal changes over 20 years. Diabetes Care 2007;
30:8616.
27. Fritsch M, Schober E, Rami-Merhar B, et al. Diabetic ketoacidosis at diagnosis in
Austrian children: a population-based analysis, 1989-2011. J Pediatr 2013;163:
14848.e1.
28. Claessen FM, Donaghue K, Craig M. Consistently high incidence of diabetic
ketoacidosis in children with newly diagnosed type 1 diabetes. Med J Aust
2012;197:216.
29. Jefferies CA, Cutfield SW, Derraik JGB, et al. 15-year incidence of diabetic ketoacidosis at onset of type 1 diabetes in children from a regional setting (Auckland,
New Zealand). Sci Rep 2015;5:10358.
30. Rewers A, Klingensmith G, Davis C, et al. Presence of diabetic ketoacidosis at
diagnosis of diabetes mellitus in youth: the Search for Diabetes in Youth Study.
Pediatrics 2008;121(5):e125866.
31. Abdul-Rasoul M, Al-Mahdi M, Al-Qattan H, et al. Ketoacidosis at presentation of
type 1 diabetes in children in Kuwait: frequency and clinical characteristics.
Pediatr Diabetes 2010;11(5):3516.
32. Komulainen J, Kulmala P, Savola K, et al. Clinical, autoimmune, and genetic characteristics of very young children with type 1 diabetes. Childhood Diabetes in
Finland (DiMe) Study Group. Diabetes Care 1999;22(12):19505.
33. Sundaram PC, Day E, Kirk JM. Delayed diagnosis in type 1 diabetes mellitus.
Arch Dis Child 2009;94(2):1512.
34. Bui H, To T, Stein R, et al. Is diabetic ketoacidosis at disease onset a result of
missed diagnosis? J Pediatr 2010;156(3):4727.
35. Cengiz E, Xing D, Wong JC, et al. Severe hypoglycemia and diabetic ketoacidosis among youth with type 1 diabetes in the T1DM Exchange clinic registry.
Pediatr Diabetes 2013;14(6):44754.
36. Rewers A, Chase HP, Mackenzie T, et al. Predictors of acute complications in children with type 1 diabetes. JAMA 2002;287(19):25118.
37. King BR, Howard NJ, Verge CF, et al. A diabetes awareness campaign prevents
diabetic ketoacidosis in children at their initial presentation with type 1 diabetes.
Pediatr Diabetes 2012;13:64751.
38. Vanelli M, Chiari G, Ghizzoni L, et al. Effectiveness of a prevention program for
diabetic ketoacidosis in children. An 8-year study in schools and private practices. Diabetes Care 1999;22:79.
39. Lansdown AJ, Barton J, Warner J, et al. Prevalence of ketoacidosis at diagnosis
of childhood onset Type 1 diabetes in Wales from 1991 to 2009 and effect of a
publicity campaign. Diabet Med 2012;29:15069.
40. Morris AD, Boyle DI, McMahon AD, et al. Adherence to insulin treatment, glycaemic control, and ketoacidosis in insulin-dependent diabetes mellitus. The DARTS/
MEMO Collaboration. Diabetes Audit and Research in Tayside Scotland. Medicines Monitoring Unit. Lancet 1997;350:150510.
41. Ellis D, Naar-King S, Templin T, et al. Multisystemic therapy for adolescents with
poorly controlled type 1 diabetes: reduced diabetic ketoacidosis admissions and
related costs over 24 months. Diabetes Care 2008;31:17467.
42. Estrada CL, Danielson KK, Drum ML, et al. Hospitalization subsequent to diagnosis
in young patients with diabetes in Chicago, Illinois. Pediatrics 2009;124(3):92634.
43. Fritsch M, Rosenbauer J, Schober E, et al. Predictors of diabetic ketoacidosis in
children and adolescents with type 1 diabetes. Experience from a large multicentre database. Pediatr Diabetes 2011;12:30712.

869

870

Jefferies et al

44. Hanas R, Lindgren F, Lindblad B. A 2-yr national population study of pediatric

ketoacidosis in Sweden: predisposing conditions and insulin pump use. Pediatr
Diabetes 2009;10(1):337.
45. Icks A, Rosenbauer J, Haastert B, et al. Hospitalization among diabetic children
and adolescents and non-diabetic control subjects: a prospective populationbased study. Diabetologia 2001;44(Suppl 3):B8792.
46. Liberatore R Jr, Perlman K, Buccino J, et al. Continuous subcutaneous insulin
infusion pump treatment in children with type 1 diabetes mellitus. J Pediatr Endocrinol Metab 2004;17(2):2236.
47. American Psychiatric Association. Diagnostic and Statistical Manual of Mental
Disorders. 4th edition. Washington, DC; 2000.
48. Jones JM, Lawson ML, Daneman D, et al. Eating disorders in adolescent females
with and without type 1 diabetes: cross sectional study. BMJ 2000;320(7249):
15636.
49. Daneman D, Olmsted M, Rydall A, et al. Eating disorders in young women
with type 1 diabetes. Prevalence, problems and prevention. Horm Res 1998;
50(Suppl 1):7986.
50. Garrison MM, Katon WJ, Richardson LP. The impact of psychiatric comorbidities
on readmissions for diabetes in youth. Diabetes Care 2005;28(9):21504.
51. Stewart SM, Rao U, Emslie GJ, et al. Depressive symptoms predict hospitalization for adolescents with type 1 diabetes mellitus. Pediatrics 2005;115(5):
13159.
52. Geffken GR, Heather L, Walker KN, et al. Family functioning processes and diabetic ketoacidosis in youths with type I diabetes. Rehabil Psychol 2008;53(2):
2317.
53. Hampson SE, Skinner TC, Hart J, et al. Behavioral interventions for adolescents with type 1 diabetes: how effective are they? Diabetes Care 2000;23:
141622.
54. Kovacs M, Charron-Prochownik D, Obrosky DS. A longitudinal study of biomedical and psychosocial predictors of multiple hospitalizations among young people with insulin-dependent diabetes mellitus. Diabet Med 1995;12:1428.
55. Nakhla M, Daneman D, To T, et al. Transition to adult care for youths with diabetes
mellitus: findings from a Universal Health Care System. Pediatrics 2009;124(6):
e113441.
56. Dumont RH, Jacobson AM, Cole C, et al. Psychosocial predictors of acute complications of diabetes in youth. Diabet Med 1995;12:6128.
57. Curtis JR, To T, Muirhead S, et al. Recent trends in hospitalization for diabetic ketoacidosis in Ontario children. Diabetes Care 2002;25:15916.
58. Randall L, Begovic J, Hudson M, et al. Recurrent diabetic ketoacidosis in innercity minority patients: behavioral, socioeconomic, and psychosocial factors.
Diabetes Care 2011;34:18916.
59. Harris MA, Harris BS, Mertlich D. Brief report: in-home family therapy for adolescents with poorly controlled diabetes: failure to maintain benefits at 6-month
follow-up. J Pediatr Psychol 2005;30:6838.
60. Wysocki T, Harris MA, Buckloh LM, et al. Randomized trial of behavioral family
systems therapy for diabetes: maintenance of effects on diabetes outcomes in
adolescents. Diabetes Care 2007;30:55560.
61. Nunn E, King B, Smart C, et al. A randomized controlled trial of telephone calls to
young patients with poorly controlled type 1 diabetes. Pediatr Diabetes 2006;7:
2549.

Preventing Diabetic Ketoacidosis

62. Couper JJ, Taylor J, Fotheringham MJ, et al. Failure to maintain the benefits of
home-based intervention in adolescents with poorly controlled type 1 diabetes.
Diabetes Care 1999;22:19337.
63. Ellis DA, Naar-King S, Chen X, et al. Multisystemic therapy compared to telephone support for youth with poorly controlled diabetes: findings from a randomized controlled trial. Ann Behav Med 2012;44:20715.
64. Grey M, Boland EA, Davidson M, et al. Coping skills training for youth with diabetes mellitus has long-lasting effects on metabolic control and quality of life.
J Pediatr 2000;137:10713.
65. Golden MP, Herrold AJ, Orr DP. An approach to prevention of recurrent diabetic
ketoacidosis in the pediatric population. J Pediatr 1985;107(2):195200.
66. Nguyen TM, Mason KJ, Sanders CG, et al. Targeting blood glucose management
in school improves glycemic control in children with poorly controlled type 1 diabetes mellitus. J Pediatr 2008;153(4):5758.

871